The nightly news brings stories about smart programs that are beating the best human players at chess or on the game show Jeopardy. As challenging as these games are, isn’t this an example of super powerful computers or fast access to mountains of databases?
Scientists and cutting edge technology firms around the world are going beyond using faster computers and bigger databases. They are creating computers that can learn and even teach themselves without human help. Computers teaching other computers or learning on their own have sped up the kinds of tasks they can perform. We are starting to see smart computers joining in activities that seem to need human judgement and nuance or making assumptions and surprisingly insightful leaps of faith. A prime example is Google’s demonstration of a smart assistant calling a salon for a hair appointment or talk to a restaurant host to make a reservation. We may not realize that humans and human behavior is a miracle, in part, because of how complex it is to copy. Think of the give and take of natural, human conversation. We adjust to help keep things going. We pause, use our tone, volume, word choice, add emotion–all communicating what we mean, what we want, and how we feel. And we react to other people in the conversation without realizing how phenomenal this is. The phone calls the Google assistant made show a limited set of skills right now but, as an example of jumping into real human conversation, this is an amazing breakthrough for virtual agents.
Most people assume all of this goes on only at the tech giants and academic researchers at elite universities. This just isn’t true. We don’t have to wait for this to evolve into something we can use in our current lives.
We all have started using these advanced tools already without even realizing it. On your email, the annoying spam never gets into your mailbox because of intelligent, adaptable filtering, and if you turn on captions on your favorite tv programs or youtube clips, many are produced or improved using artificial intelligence similar to the Google assistant that made the realistic phone calls.
The newest smart inventions coming into our daily lives are the virtual agents that are becoming popular. From the talking virtual agents like Apple’s Siri, Amazon’s Alexa and Google’s Assistant, to chatty helpers that pop up on websites and inside apps, everyone is able to just talk to get help and answer their deepest questions any time of day. Now it is becoming common for websites to have virtual agents–human language programs that pop up on websites and have conversations with customers. Banks were the first to go all in on this technology, which is not surprising since they have been keen to reduce their labor costs by adopting online and automated services.
In the insurance world, both carriers and agencies alike are also launching these tools with smart automation. The first smart virtual agents were able to take what the customer typed and search for suggested answers, then these virtual agents progressed, with friendly scripts and the smarts that made them seem more human. This previous generation of insurance virtual agents were limited in what they could talk about. Now we are starting to see some really surprising tools–automated smart agents–that can take care of tasks from start to finish.
Right now the virtual agent products are highly structured in what information they can offer, tasks they can perform, and what they can say. We provide the structure and the data these helpers deliver to our customers. Where do the smarts come in? This generation of intelligent virtual agents are trained to understand the vast variety of ways we humans phrase things. Think about how many ways a customer can ask for proof of insurance or about their insurance bill? The smart virtual insurance agents have been trained with millions of phrases that translate to the hundreds of things the customer really intended to ask. The brain of the virtual agent doesn’t just match the request to one it has learned before. It can recognize requests that are pretty close to a match and add this to its previous learning. Just like a toddler, the virtual agents listen, get feedback and add to their vocabulary and understanding.
In an Elafris survey of agency call center issues, it found one-third of the support requests concerned billing and payments, with another twenty-percent being calls about quotes and policy questions. All of these are easily handled by well-trained virtual insurance agents. 53% of service request traffic can be eliminated instantly from your live agent queues. More than cutting down on your live agents’ workload, it enables CSRs to focus on being friendly and helpful, building customer loyalty, wait times for live agents are reduced and during off-hours, virtual agents can take care of the majority of your customer’s needs. All of this makes for happier policyholders and prospective customers.
Intelligent virtual agents are great students. They are being schooled in almost every area of the insurance industry, even complex processes like claims, or niche insurance products. Right now virtual agents are learning how to manage claims from start to finish. Imagine a technology that helps policyholders with all of the documents, photos, estimates and updates, right through to sending the payment to the repair shops and the customer. The same technology can analyze claims information against similar claims to the company to potential fraud or other risks.
For niche insurance products, it can be difficult to justify marketing and servicing them well. With automation of the quoting and billing, suddenly lines that were lagging in revenue can start carrying their weight. Integrate your niche set of products with a well-targeted online ads, and these smart insurance agents can transform into winning producers for the smaller product lines.
Virtual agents make sense for carriers and agencies right now to help solve the most pressing needs. To see how Elafris Virtual Insurance Agents can help your company click here to schedule a free online demo.